Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Giuseppe Bianchi
bianchi@elet.polimi.it
Giuseppe Bianchi
Course outline
ÎPart 2: GSM
ÎPart 3: Wi-Fi
ÎExtra time?
Giuseppe Bianchi
1
Wireless communication
ÎEarly wireless communication:
in the 400-900 TeraHertz Band!
Ö150 BC smoke signals (Greece)
Ö1794, optical telegraph, Claude Chappe
Giuseppe Bianchi
types of communication
ÎSimplex
Öone-way communication
Æradio, TV, etc
ÎHalf-duplex:
Ötwo-way communication but not simultaneous
Æpush-to-talk radios, etc
ÎFull-duplex:
Ötwo-way communication
Æcellular phones
ÖFrequency-division duplex (FDD)
ÖTime-division duplex (TDD): simulated full-duplex
Giuseppe Bianchi
2
Why wireless communication?
ÎUser Mobility
ÎReduced Cost (cheap infrastructure)
ÖCabling very critical
ÖDeveloping nations utilize cellular telephony rather than
laying twisted-pair wires to each home
ÎFlexibility
ÖCan easily set-up temporary LANs
ÆDisaster situations
ÆOffice moves
ÎOnly use resources when sending or
receiving a signal
Giuseppe Bianchi
Giuseppe Bianchi
3
History of wireless
communication
Î 1896: Marconi
Ö first demonstration of wireless telegraphy
Ö tx of radio waves to a ship at sea 29 km away
Ö long wave transmission, high power req. (200 kW and +)
Î 1901: Marconi
Ö Telegraph across the atlantic ocean
Ö Close to 3000 Km hop!
Î 1907 Commercial transatlantic connections
Ö huge ground stations (30 by100m antennas)
Î 1915: Wireless telephony established
Ö NY – S. Francisco
Ö Virginia and Paris
Î 1920 Marconi:
Ö Discovery of short waves (< 100m)
Ö reflection at the ionosphere
Ö (cheaper) smaller sender and receiver, possible due to the invention of the
vacuum tube (1906, Lee DeForest and Robert von Lieben)
Giuseppe Bianchi
History of wireless
communication
Î 1920's: Radio broadcasting became popular
Î 1928: many TV broadcast trials
Î 1930's: TV broadcasting deployment
Î 1946: First public mobile telephone service in US
Ö St. Louis, Missouri
Ö Single cell system
Î 1960's: Bell Labs developed cellular concept
Ö brought mobile telephony to masses
Î 1960’s: Communications satellites launched
Î Late 1970's: technology advances enable
affordable cellular telephony
Ö entering the modern cellular era
Î 1974-1978: First field Trial for Cellular System
Ö AMPS, Chicago
Giuseppe Bianchi
4
1st generation mobile systems
ÎFirst generation: 1980’s ÎAnalog transmission
ÎSeveral competing standards in Ö Frequency modulation
different countries ÎVarious bands:
Ö NMT (Nordic Mobile Telephone) Ö NMT:
ÆScandinavian standard; adopted in most Æ450 MHz first
of Europe Æ900 MHz later
ÆFirst european system (Sweden, 1981) Ö TACS
Ö TACS (Total Access Communication Æ900 MHz
Systems), starts in 1985 Ö AMPS
ÆUK standard; A few of Europe, Asia,
Æ800 MHz
Japan
Ö AMPS (Advanced Mobile Phone ÎToday still in use in low-
Service) technology countries
ÆUS standard Ö And not yet completely
Ö C-Netz (Only in Germany) dismissed in high-tech
Ö Radiocom 2000 (Only in France) countries
Giuseppe Bianchi
Giuseppe Bianchi
5
Timing
Î1982: Start of GSM-specification in
Europe (1982-1990)
Î1983: Start of American AMPS
widespread deployment
Î1984 CT-1 standard (Europe) for
cordless telephones
Î1991 Specification of DECT
ÖDigital European Cordless Telephone (today: Digital
Enhanced Cordless Telecommunications)
- ~100-500m range, 120 duplex channels, 1.2Mbit/s data
transmission, voice encryption, authentication
Î1992: Start of GSM operation Europe-
wide
Î1994: DCS-1800
Giuseppe Bianchi
Giuseppe Bianchi
6
3rd generation mobile systems
Giuseppe Bianchi
Giuseppe Bianchi
7
WLAN speeds
Giuseppe Bianchi
Giuseppe Bianchi
8
WLAN Market - HotSpots
U.S. Commercial Hotspots
2001-2002: exceeded
U.S. Hotspots growth
expectations by 14%
(2002-2006)
4.500 50000
4.000 40000
3.500
30000
3.000
20000
2.500
10000
2.000
1.500 0
2002 2004 2006
2001 2002
Unique U.S. Hotspots
Forecasted Actual
Giuseppe Bianchi
9
PART 1
Propagation Characteristics of
Wireless Channels
Lecture 1.1
Basic concepts and
terminology
Giuseppe Bianchi
f
f = 900 MHz Æ λ = 33 cm
Giuseppe Bianchi
1
The radio spectrum
ELF <3 KHz Remote control, Voice, analog phone
VLF 3-30 KHz Submarine, long-range
LF 30-300 KHz Long-range, marine beacon
MF 300 KHz –3 MHz AM radio, marine radio
HF 3-30 MHz Amateur radio, military, long-distance aircraft/ships
VHF 30-300 MHZ TV VHF, FM radio, AM x aircraft commun.
UHF 300 MHz - 3 GHz Cellular, TV UHF, radar
SHF 3-30 GHz Satellite, radar, terrestrial wireless links, WLL
EHF 30-300 GHz Experimental, WLL
IR 300 GHz – 400 THz LAN infrared, consumer electronics
Light 400-900 THz Optical communications
Giuseppe Bianchi
Impairments due to
-Oxygen
- water vapour
Giuseppe Bianchi
2
Spectrum Allocation
Î Cellular systems
Ö 400-2200 MHz range (VHF-UHF)
Ö Simple, small antenna (few cm)
Ö With less than 1W transmit power, can cover several floors within
a building or several miles outside
Î SHF and higher for directed radio links,
satellite communication
Ö Large bandwidth available
Î wireless data systems
Ö 2.4, 5 GHz zones (ISN band)
Ö Main interference from microwave ovens
Ö limitations due to absorption by water and oxygen - weather
dependent fading, signal loss due to by heavy rainfall etc.
Giuseppe Bianchi
Higher-lower Frequencies
ÎHigher frequencies:
Ömore bandwidth
Öless crowded spectrum
Öbut greater attenuation through walls
ÎLower frequencies
Öbandwidth limited
Ölonger antennas required
Ögreater antenna separation required
Öseveral sources of man-made noise
Giuseppe Bianchi
3
Antennas
Giuseppe Bianchi
Antenna Gain
Î Isotropic antenna (idealized)
Ö Radiates power equally in all directions (3D)
Ö Real antennas always have directive effects (vertically and/or
horizontally)
Î Antenna gain
Ö Power output, in a particular direction, compared to that produced in any
direction by a perfect omni-directional antenna (isotropic antenna)
power density at a distance d in the direction of maximum radiation
Directivity D =
mean power density at a distance d
power density at a distance d in the direction of maximum radiation
Gain G =
PT / 4πd 2
4
Radiation Patterns
Î Graphical representation of radiation properties
of an antenna
Î Depicted as two-dimensional cross section
y y z
simple
x z x dipole
side view (xy-plane) side view (yz-plane) top view (xz-plane)
y y z
directed
x z x antenna
Giuseppe Bianchi
z
z
x
sectorized
x antenna
Giuseppe Bianchi
5
Three means of propagation
Giuseppe Bianchi
6
Ionospheric or Sky Wave
Propagation
Line-of-Sight Propagation
7
Radio horizon
Optical
horizon Radio
horizon
optical horizon : d = 3.57 h
radio horizon : d = 3.57 K ⋅ h K ≈ 4/3
d = [ Km], h = [m]
Giuseppe Bianchi
Giuseppe Bianchi
8
Propagation impairments
ÎLine of sight
ÎReflection
ÎShadowing
Giuseppe Bianchi
Propagation impairments
BS
ÎDiffraction MS
Î When the surface
encountered has sharp edges
Æ bending the wave
BS
ÎScattering
Î When the wave encounters
objects smaller than the
wavelength (vegetation,
clouds, street signs)
Giuseppe Bianchi
9
Signal attenuation
Signal power
Distance BS Æ MS
Giuseppe Bianchi
Multipath Characteristics
(not just attenuation)
10
Slow fading – fast fading
Signal power
Distance BS Æ MS (m)
Distance BS Æ MS (km)
Giuseppe Bianchi
Time
Giuseppe Bianchi
11
PART 1
Propagation Characteristics of
Wireless Channels
Lecture 1.2
Attenuation models
Giuseppe Bianchi
10 log(P1 / P2 )
PA = 1 Watt
PB = 50 milliWatt
Æ PA = 13 dB greater than PB
Giuseppe Bianchi
12
Decibels - dBm
ÎdBm = absolute value (reference= 1mW)
ÖPower in dBm = 10 log(power/1mW)
ÖPower in dBW = 10 log(power/1W)
» Not much used by us 1dBW=30dBm
ÎExamples
Ö10 mW = 10 log10(0.01/0.001) = 10 dBm
Ö10 µW = 10 log10(0.00001/0.001) = -20 dBm
Ö26 dBm = ___ 2W= ___ dBm?
ÖS/N ratio = -3dB Æ S = ___ X N?
ÎProperties & conversions
ÖdBm = 10 log10(P (W) / 1 mW) = P (dB) + 30 dBm
ÖP1 * P2 (dBm) = P1 (dBm) + P2 (dB)
P1 * P2 (dBm) = 10 log10(P1*P2 (W)/0.001) =
10log10(P1/0.001) + 10 log10P2 = P1 (dBm) + P2 (dB)
Giuseppe Bianchi
Computation with dB
ÎTransmit power
ÖMeasured in dBm
ÆEs. 33 dBm
ÎReceive Power
ÖMeasured in dBm
ÆEs. –10 dBm
ÎPath Loss
ÖReceive power / transmit power
ÖMeasured in dB
ÖLoss (dB) = transmit (dBm) – receive (dBm)
ÆEs. 43 dB = attenuation by factor 20.000
Giuseppe Bianchi
13
Attenuation model for LOS
ÎDirect path between transmitter and
receiver
Öunobstructed Line-Of-Sight (LOS)
ÖRadio signal behaves like light in free space (straight line)
ÎReceive power:
ÖIn absence of obstacles, received power follows inverse
square law
Pr (d ) ∝ d −2
Æ(d = distance between sender and receiver)
Giuseppe Bianchi
14
free space model – real antennas
r
ÎNon isotropic tx antenna
ÖAntenna gain Gt
Gt Pt
Pa (d ) =
4πd 2
Giuseppe Bianchi
Î Pt = transmitter power
Î Pt Gt = Equivalent Isotropic
Ö (W or mW) Radiated Power (EIRP)
Î Gt = transmitter antenna gain Î L = other system losses (hardware)
Î Gr = transmitter antenna gain Ö L >=1 (dimensionless)
Ö (dimensionless) Î d = distance between transmitter
Î λ = c/f = RF wavelength (m) and receiver (m)
Ö c = speed of light (3x108 m/s)
Ö f = RF frequency (Hz)
15
Example
normalized
frequency [MHz] 900 900000000
speed of light [Km 300000 300000000
lambda (m) 0,333333333
gain Tx 1
Gain Rx 1
Loss 1
Ptx [W] 5
distance (Km) Prx W Prx dBm
200 8,80E-08 -40,56
400 2,20E-08 -46,58
600 9,77E-09 -50,10 -30,00
800 5,50E-09 -52,60
1000 3,52E-09 -54,54
Giuseppe Bianchi
P ⎧⎪ L ⎛ 4πd ⎞
2
⎫⎪
PL(d )[dB ] = 10 log10 t = 10 log10 ⎨ ⎜ ⎟ ⎬=
Pr ⎪⎩ Gt Gr ⎝ λ ⎠ ⎪⎭
Gt Gr λ
= 20 log10 d − 10 log10 − 20 log10 =
L 4π
GG c
= 20 log10 d + 20 log10 f − 10 log10 t r − 20 log10 =
L 4π
GG
= 20 log10 d + 20 log10 f − 10 log10 t r − 147.56
L
Giuseppe Bianchi
16
Free space loss
same as path loss, but part due to attenuation
in free space only (in dB)
−2
⎛ λ ⎞
L free (d ) = ⎜ ⎟
⎝ 4πd ⎠
⎡ λ ⎤ ⎡c / f ⎤
L free (d ) [dB ]= −20 log ⎢ = −20 log
⎣ 4πd ⎥⎦ ⎢⎣ 4πd ⎥⎦
= 20 log10 d + 20 log10 f − 147.56
Giuseppe Bianchi
Reference distance
Î If known received power at a reference distance do
from tx 2
Ö can calculate Pr(d) for any d ⎛ do ⎞
Pr (d ) = Pr (d o )⎜ ⎟
⎝d ⎠
⎛d ⎞ ⎛d ⎞
Pr (d )(dBm) = 10 log10 Pr (d o ) + 20 log10 ⎜ o ⎟ = Pr (d o )(dBm) + 20 log10 ⎜ o ⎟
⎝d ⎠ ⎝d ⎠
Giuseppe Bianchi
17
Example
18
More realistic propagation
models
ÎInverse square power law
ÖWay too optimistic (ideal case)
ÖReal world: η-th power law
Pr (d ) ∝ d −η
Öwith η ranging up to as much as η=7
ÆIf tough environment (e.g., lots of foliage),
Ötypical values:
Æη=2 for small distances (20 dB/decade)
Æη=3 to η=4 (40 dB/decade) for mobile telephone distances
Öη higher in cities and urban areas; η lower in suburban or
rural areas.
Giuseppe Bianchi
Extended formulae
⎛d ⎞
Pr (d )(dB ) = 10 log10 Pr (d o ) + 10η log10 ⎜ o ⎟
⎝d ⎠
d_ref 1 Km
P_ref -51,5266 dBm (Ptx=10W; 900 MHz; 1000m)
Giuseppe Bianchi
19
Realistic scenarios
Î obstructions between the transmitter and
receiver
Ö reflection, diffraction, scattering
Ö Propagation strongly influenced by environment (building characteristics,
vegetation density, terrain variation)
Ö Perfect conductors reflect waves; nonconductors absorb some energy
Î wave traverses multiple paths
Ö Radio waves arrive at receiver from different directions and with different
time delays
Î Resultant signal at receiving antenna is vector
addition of incoming signals
Ö signals can add constructively (resultant signal has large power) or
destructively (resultant signal has small power) depending on relative
phases
Example scenarios:
LOS path non necessarily existing (and unique)
Example: city with large buildings;
No LINE OF SIGHT;
Diffraction; reflection
diffraction
reflection
Giuseppe Bianchi
20
Example scenarios
LINE OF SIGHT +
Diffraction, reflection, scattering
LOS
Giuseppe Bianchi
Two-Ray Ground
Propagation Model
ÎTheoretical foundation for η=4
ÖTwo-ray model assumes one direct LOS path and one
reflection path reach receiver with significant power
ÖEasy to solve
Line-Of-Sight ray
ht
hr
reflected ray
Giuseppe Bianchi
21
Two-ray model – geometry
d >> ht , hr
ddirect
ht
hr
θ( )θ
dreflect
1/ 2
⎧⎪ ⎛ h − h ⎞ 2 ⎫⎪ ⎧⎪ 1 ⎛ h − h ⎞ 2 ⎫⎪
d direct = d + (ht − hr ) = d ⎨1 + ⎜ t r ⎟ ⎬ ≈ d ⎨1 + ⎜ t r ⎟ ⎬
2 2
⎪⎩ ⎝ d ⎠ ⎪⎭ ⎪⎩ 2 ⎝ d ⎠ ⎪⎭
1/ 2
⎧⎪ ⎛ ht + hr ⎞ 2 ⎫⎪ ⎧⎪ 1 ⎛ ht + hr ⎞ 2 ⎫⎪
d reflect = d + (ht + hr ) = d ⎨1 + ⎜ ⎟ ⎬ ≈ d ⎨1 + ⎜ ⎟ ⎬
2 2
⎪⎩ ⎝ d ⎠ ⎪⎭ ⎪⎩ 2 ⎝ d ⎠ ⎪⎭
Giuseppe Bianchi
Giuseppe Bianchi
22
Two ray model – field strength
ÎPhase difference
∆d 2π 4πht hr
∆ϕ = 2πf = ∆d =
c λ λd
ÎReceived field strength
ÖLet Edirect be the field strength given by direct ray.
ÖThen E = E [
1 + ρe − j∆ϕ
direct ]
ÎAssume ideal reflection (ρ=-1)
[ ]
E = Edirect 1 − e − j∆ϕ = Edirect [1 − cos ∆ϕ + j sin ∆ϕ ]
[
E = Edirect 1 + cos 2 ∆ϕ − 2 cos ∆ϕ + sin 2 ∆ϕ ]
1/ 2
=
1 − cos ∆ϕ ∆ϕ
= 2 Edirect = 2 Edirect sin
2 2
Giuseppe Bianchi
PG G ⎛ λ ⎞ 2 ⎛ 2πht hr ⎞
2
Pr (d ) = t t r ⋅ ⎜ ⎟ ⋅ 4 sin ⎜ ⎟
L ⎝ 4πd ⎠ ⎝ λd ⎠
Giuseppe Bianchi
23
Two ray model - conclusion
Î Typical values:
Ö ht ~ few tens of m 2πht hr ⎛ 2πht hr ⎞ ⎛ 2πht hr ⎞
2
Pt Gt Gr ⎛ λ ⎞ ⎛ 2πht hr ⎞
2 2
Pt Gt Gr ht2 hr2
Pr (d ) ≈ ⋅⎜ ⎟ ⋅ 4⎜ ⎟ = ⋅ 4
L ⎝ 4πd ⎠ ⎝ λd ⎠ L d
Giuseppe Bianchi
Design notes
Giuseppe Bianchi
24
Empirical models
ÎConsider specific scenarios
ÖUrban area (large-medium-small city), rural area
ÖModels generated by combining most likely ray traces
(LOS, reflected, diffracted, scattered)
ÖBased on large amount of empirical measurements
ÎAccount for parameters
ÖFrequency; antenna heights; distance
ÎAccount for correction factors
Ö(diffraction due to mountains, lakes, road shapes, hills, etc)
First model: Okumura, 1968
VERY complex due to many specific correction factors!
Giuseppe Bianchi
Okumura-Hata model
ÎHata (1980): very simple model to fit
Okumura results
ÎProvide formulas to evaluate path loss
versus distance for various scenarios
ÖLarge cities; Small and medium cities; Rural areas
ÖLimit: d>=1km
Parameters:
Æf = carrier frequency (MHz) Effective BS
Antenna height
Æd = distance BS Æ MS (Km)
Æhbs = (effective) heigh of base
station antenna (m)
Æhms = height of mobile antenna (m)
Giuseppe Bianchi
25
Okumura-Hata: urban area
L path (dB ) = 69.55 + 26.16 log10 f +
+ (44.9 − 6.55 log10 hbs ) log10 d +
− 13.82 log10 hbs − a(hms )
Î a(hms) = correction factor to differentiate large from
medium-small cities;
Î depends on MS antenna height
small - med cities : a(hms ) = [1.1 log10 f − 0.7]hms − [1.56 log10 f − 0.8]
Very small correction difference between large and small cities (about 1 dB)
Giuseppe Bianchi
2
⎡ f ⎤
suburban : L path (dB ) = L p − 2 ⎢log10 ⎥ − 5.4
⎣ 28 ⎦
L path ( dB) = L p − 4.78[log10 f ] + 18.33 log10 f − 40.94
2
rural :
Giuseppe Bianchi
26
Okumura-Hata: examples
150
140
130
path loss (dB)
120
110
large cities
100 small cities
suburbs
90
rural area
80
0 1 2 3 4 5 6 7 8 9 10
distance (km)
Giuseppe Bianchi
Okumura-Hata and η
Î Coefficient of Log(d) depends only on hbs
Î 10η = attenuation (dB) in a decade
Ö (d=1 Æ d=10)
Î The higher the BS, the lower the coefficient η
44
42
40
38
10η
36
34
32
30
0 20 40 60 80 100 120 140
base station height (m)
Giuseppe Bianchi
27
Other empirical models
Î Lee’s model
Ö Use at 900MHZ
Ö For distances > 1km
Ö Based on measurements taken in three cities (including Philadelpia)
Ö More complex than Okumura-Hata
Î Walfish-Ikegami model
Ö For frequency range 800-2000 MHz
Ö Valid for microcellular distances (20m – 5 km)
Ö Adopted by European Cooperation in the field of Scientific and Technical
(COST) research as reference model for 3G systems
Î Indoor propagation models
Ö Include attenuation factors due to building penetration
Ö Account for number of walls, floors, reflection loss, etc
Ö Based on zones (large zone, middle zone, small zone, microzone)
Giuseppe Bianchi
PART 1
Propagation Characteristics of
Wireless Channels
Lecture 1.3
fading models
Giuseppe Bianchi
28
Statistical nature of received
power
Signal strength (dB)
Mean value
predicted
by attenuation
model (constant
Long term fading at given )
Giuseppe Bianchi
er (t ) = ∑k =1 ak cos(2πf 0t + φ k )
N
29
Multipath analysis
er (t ) = ∑k =1 ak cos(2πf 0t + φ k ) =
N
recall that : cos(2πf 0t + φ k ) =
= cos(2πf 0t ) cos(φ k ) − sin (2πf 0t )sin (φ k )
Rayleigh distribution
0,7
= Pr ( x ≤ a < x + dx ) =
0,3
0,2
2
x − x 2σ 2
0,1
= 2e
0 σ
0 1 2 3 4 5
amplitude
∞
π
2
E [a ] = ∫ x ⋅
x −x
e 2σ 2
=σ = 1.253σ
0
σ 2
2
∞
π ⎛ π⎞
2
Var [a ] = ∫ x 2 ⋅
x −x
e 2σ 2
−σ 2 = σ 2 ⎜ 2 − ⎟ = 0.4292σ 2
0
σ 2
2 ⎝ 2⎠
Giuseppe Bianchi
30
Signal power
amplitude : a = X 2 + Y 2 : rayleigh distribution
power : p = a 2 = X 2 + Y 2 : exponential distribution
ÎAverage power:
Ö 2σ2 (average over time)
ÎInstantaneous power
ÖRandom variable, with: −x
f p (x ) =
1 2σ 2
Æprobability density function: e
2σ x 2
−
Fp ( x ) = 1 − e 2σ
2
ÆProbability distribution function:
Giuseppe Bianchi
Outage probability
ÎProbability that received power is lower
than a given threshold
ÖBelow which signal cannot be correctly received
Example 1: Example 2:
average power=100 µW; average power= -13 dBm;
lower threshold=15µW; lower threshold=-30 dBm;
Outage probability= 1-exp(-15/100) = 13,9% Outage probability= 1-exp(-1/50) = 1,98%
Giuseppe Bianchi
31
Long-term fading
Giuseppe Bianchi
Giuseppe Bianchi
32
Long-term fading and
attenuation plot
attenuation: η=4 after 100m; η=2 before 100m
-30
-40
no shadowing
-50 sigma=3 dB
received power (dBm)
sigma=8 dB
-60
-70
-80
-90
-100
-110
-120
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
distance (m)
Giuseppe Bianchi
33
normal distribution
0.4
σ=1
0.3
0.2
σ=2 Standard:
µ=0; σ=1
0.1
σ=3
-6 -4 -2 µ=0 2 4 6
− ( x − µ )2
1
f X ( x) = e 2σ 2
2π ⋅ σ
Giuseppe Bianchi
1
µ=0; σ=1
0.8
0.6
0.4
0.2
-4 -2 2 4
x t2
−
Q( x ) = FX ( x) =
1
2π ∫e
−∞
2
dt
Giuseppe Bianchi
34
erf - erfc
x
erf ( x )=
2
∫ dt
−t 2
e
π 0
erfc( x) = 1 − erf ( x)
t2
x x x −
2⋅(1 / 2 )
erf ( x )=
2 1 1
∫ e dt = ∫ e dt = ∫
2
−t 2
−t2
π 0 π −x (
2π ⋅ 1 / 2 ) −x
e dt
Giuseppe Bianchi
⎛ x ⎞ 1
erfc⎜ ⎟
⎝ 2 ⎠ 0.8
2
0.6
0.4
0.2
-4 -2 2 4
Giuseppe Bianchi
35
Normal distribution (standard)
-4 0,003%
-3,9 0,005%
-3,8 0,007%
-3,7 0,011%
100,000%
-3,6 0,016%
-3,5 0,023%
-3,4 0,034%
-3,3 0,048% 10,000%
-3,2 0,069%
-3,1 0,097%
-3 0,135%
-2,9 0,187% outage prob 1,000%
-2,8 0,256%
-2,7 0,347%
-2,6 0,466%
-2,5 0,621% 0,100%
-2,4 0,820%
-2,3 1,072%
-2,2 1,390%
-2,1 1,786% 0,010%
-2 2,275%
-1,9 2,872%
-1,8 3,593%
-1,7 4,457% 0,001%
-1,6 5,480%
-1,5 6,681% -4 -3 -2 -1 0
-1,4 8,076%
-1,3 9,680% n x sigma
-1,2 11,507%
-1,1 13,567%
-1 15,866%
Giuseppe Bianchi
PART 1
Propagation Characteristics of
Wireless Channels
Lecture 1.4
coverage area estimation
(cell sizing)
Giuseppe Bianchi
36
Cell radius
ÎOpen issue, until now:
ÖHow do we determine the cell radius?
Example (part 1)
37
Fading Margin
Î Previous computation does
not account for long-term
fading
Ö Need to keep it in count, as it does not
reduce when the MS makes small moves L p = Pt − Pth
Ö IDEA: reduce cell radius to account for
a “fading margin” M
L p = Pt − Pth − M
Example (part 2)
38
Outage probability (1)
Recall that slow fading has lognormal distribution. In dB:
− ( pdB ( r )− Pav ,dB ( r ))2
fY ( pdB (r ))=
1 2σ dB
2
e
2π ⋅ σ dB
pdB(r)= power (dBm) received at given distance r
Pav,dB(r)= mean power (dBm, from attenuation laws) at same distance
−∞
1 ⎛ P (r ) − Pth ⎞
= erfc⎜⎜ av ,dB ⎟
2 ⎝ 2 ⋅ σ dB ⎟⎠
Giuseppe Bianchi
39
Outage probability (3)
If outage probability is specificaly computed for a terminal placed
at cell border:
⎛ M ⎞
Pout (cell border ) = erfc⎜⎜
1
⎟
2 ⋅ σ ⎟
⎝ 2 dB ⎠
Giuseppe Bianchi
Example (part 3)
Î Received power at 10 mt: 30%
0.1 W
25%
Î Threshold power:
Pth = -50 dBm
outage probability
20%
Î η = 3.7
Î Slow fading, with 15%
σdB=4 dB
Î fading margin M=6 10%
5%
Î Cell size =537mt
Î Poutage= 6,68% 0%
400 450 500 550 600 650 700
terminal distance (mt)
⎛ 70 − 37 log(r / 10) ⎞
plot : Pout (r ) = erfc⎜⎜
1
⎟
2 ⋅ σ ⎟
⎝ 2 dB ⎠
Giuseppe Bianchi
40
Outage probability area (1)
To compute the outage probability in a circular area with radius R, we just need
to integrate Pout(r) over the cell area:
⎪⎩ Pth [ mW ]
Giuseppe Bianchi
=
2
erfc ( x ) =
2 2
∫e ∫e
−t 2 −t 2
dt = 1 − dt
π π
2 2 x 0
D [erfc ( x )] = −
2
e−x
2
π
Giuseppe Bianchi
41
Outage probability area versus M
1,E+00
sigma = 4 dB
1,E-01
sigma = 6 dB Represents the fraction
sigma = 8 dB
of “not covered” area in
outage probability (area)
a cell.
1,E-02
Application example:
1,E-04
If σdB=6, and target
is 1% outage area,
1,E-05 must set M~10
0 5 10 15 20 25
fading margin M (dB)
Giuseppe Bianchi
sigma = 8 dB
(conservative)
1,E-02
Giuseppe Bianchi
42
Cell radius computation
Î Step 1:
Ö From outage probability target (es. not EXAMPLE
greater than 2%)
-Outage (border) = 5%
ÆOn a per area basis
ÆOr on a border cell basis
−σdB = 6dB
From SW (or tables)
Ö Plus radio channel information (η, σdB)
M=1.65 σdB = 9.87 dB
Ö Compute M
Î Step 2: -Acceptable performance: -90 dBm
Ö From Power threshold Pth, −η = 3.7
-Received power = -65 dBm @ 3 Km
Ö Plus transmitted power Pt
⎛d ⎞
Æor equivalent information (es. − 90 + m = −65 − 10η log10 ⎜⎜ ⎟⎟
received power at reference ⎝ d0 ⎠
distance) d~7.7 Km
Ö Compute radius
Giuseppe Bianchi
Giuseppe Bianchi
43
Slow+Fast fading margins
power
Transmitted power
Pth + M1 + M2
Short-term fading margin M2
Pth + M1
Long-term fading margin M1
Power threshold Pth = Pt - Ploss
1% - 2%
44
PART 2
Cellular Coverage Concepts
Lecture 2.1
why cells
Giuseppe Bianchi
d
1 Base Station
N=12 channels
BS •(e.g. 1 channel = 1 frequency)
Giuseppe Bianchi
1
Cellular coverage
target: cover the same area with a larger number of BSs
19 Base Station
12 frequencies
4 frequencies/cell
Worst case:
4 calls (all users in same cell)
Best case:
76 calls (4 users per cell)
Average case >> 12
Low transmit power
Key advantages:
•Increased capacity (freq. reuse)
•Decreased tx power
Giuseppe Bianchi
Disadvantage:
mobility management
Giuseppe Bianchi
2
Cellular system architecture
Î 1 BS per cell Wired network
BSs f4
f3
f2
BS = Base Station
Î MSC connected to MSC = Mobile Switching Centre
PSTN PSTN = Public Switching Telephone Network
Giuseppe Bianchi
Cellular capacity
Î Increased via frequency reuse
Ö Frequency reuse depends on interference
Ö need to sufficiently separate cells
Æreuse pattern = cluster size (7 Æ 4 Æ 3):
discussed later
Giuseppe Bianchi
3
hexagonal cells
Î Hexagon:
A D Ö Good approximation for circle
B C B
A D A
C B C
B A D A
B C B
A D A D
C B C Ö Ideal coverage pattern
D A D Æno “holes”
B C Æno cell superposition
ÎExample case:
ÖReuse pattern = 4
Giuseppe Bianchi
4
PART 2
Cellular Coverage Concepts
Lecture 2.2
Clusters and CCI
Giuseppe Bianchi
Reuse patterns D R
ÎReuse distance: 7
ÖKey concept 7 6 2
ÖIn the real world depends on 6 2
1
ÆTerritorial patterns (hills, etc) 1
5 3
5 3 4
ÆTransmitted power
4
» and other propagation issues
such as antenna directivity, Cluster: K = 7
height of transmission
antenna, etc D
ÎSimplified hexagonal 4
1
2
cells model: 3
1 1
Öreuse distance depends on 4 2 4 2
reuse pattern (cluster size) 3
1
3
ÖPossible clusters: 4 2
3
Æ3,4,7,9,12,13,16,19,…
K=4
Giuseppe Bianchi
5
Reuse distance
Î General formula D = R 3K
Î Valid for hexagonal geometry
Î D = reuse distance
Î R = cell radius
Î q = D/R =frequency reuse factor
K q=D/R
3 3,00
4 3,46
7 4,58
9 5,20
12 6,00
13 6,24
Giuseppe Bianchi
Proof
v
Î Distance between two cell
centers:
Ö (u1,v1) ÅÆ (u2,v2)
(3,2) D= [(u 2 ] [
2
− u1 ) cos 30 o + (v2 − v1 ) + (u 2 − u1 ) sin 30 o ]
2
u Ö Simplifies to:
D = (u 2 − u1 ) 2 + (v2 − v1 ) 2 + (u 2 − u1 )(v2 − v1 )
(1,1)
Ö Distance of cell (i,j) from (0,0):
30°
D = i 2 + j 2 + ij 3 R
DR = i 2 + j 2 + ij
Ö Cluster: easy to see that
K = DR2 = i 2 + j 2 + ij
Ö hence:D = R 3K
Giuseppe Bianchi
6
Clusters
Clusters:
• Number of BSs comprised in
a circle of diameter D
• Number of BSs whose inter-
distance is lower than D
K=12(i=3,j=1)
Possible clusters
all integer i,j values
i j K=ii+jj+ij q=D/R
1 0 1 1,73
1 1 3 3,00
2 0 4 3,46
2 1 7 4,58
2 2 12 6,00
3 0 9 5,20
3 1 13 6,24
3 2 19 7,55
3 3 27 9,00
4 0 16 6,93
4 1 21 7,94
4 2 28 9,17
4 3 37 10,54
4 4 48 12,00
5 0 25 8,66
5 1 31 9,64
Giuseppe Bianchi
7
Co-Channel Interference
A
G E A Î Frequency reuse implies that
C F G E remote cells interfere with
D C F tagged one
A B D C
E A B Î Co-Channel Interference (CCI)
F G E A Ö sum of interference from
C F G remote cells
B D C F
A B D
S signal power (S)
E A =
N noise power (N S ) + interferin g signal power (I)
S signal power (S)
=
I interferin g signal power (I)
S S
≈ as N S small
N I
Giuseppe Bianchi
CCI Computation -
assumptions
Î Assumptions Î Key simplification
ÖNI=6 interfering cells Ö Signal for MS at distance R
ÆNI=6: first ring interferers Ö Signal from BS interferers at
only distance D
Æwe neglect second-ring
interferers
ÖNegligible Noise NS Dint R
Power R Power
ÆS/N ~ S/I Po D Po
Öd−η propagation law Dint ~ D
Æη=4 (in general)
ÖSame parameters for all BSs
ÆSame Ptx, antenna gains, etc
Giuseppe Bianchi
8
CCI computation
S S cost ⋅ R −η By using the assumptions of
≈ = =
∑ −η same cost and same D:
NI
N I
k =1
cost ⋅ D
−η η
1 ⎛R⎞ 1 ⎛ D⎞ 1 η Results depend
= ⎜ ⎟ = ⎜ ⎟ = q on ratio q=D/R
NI ⎝ D⎠ NI ⎝R⎠ N I (q=frequency reuse factor)
(3K ) 2 = (3K )
−η
S S 1 ⎛ R ⎞ 1 η 2
≈ = ⎜ ⎟ =
N I N I ⎝ R 3K ⎠ NI 6
S (3K )
2
3
NI=6,µ=4 Æ = = K2
I 6 2
USAGE: Given an S/I target, cluster size K is obtained
Giuseppe Bianchi
Examples
Îtarget conditions: Îtarget conditions:
ÖS/I=9 dB ÖS= 18dB
Öη=4 Öη=4.2
ÎSolution: ÎSolution:
S
= 100.9 = 7.94 ≈ 8
S
[dB] = 5η log(3K ) − 10 log 6
I I
S (3K )
η
18 + 7.78
=
2
⇒K=
2 S
⋅ log(3K ) = = 1.23
I 6 3 I 21
η =4 101.23
K≥ = 5.63 ⇒ K = 7
K ≥ 2 .3 ⇒ K = 3 3
Giuseppe Bianchi
9
S/I computation
assuming 6 interferers only (first ring)
Giuseppe Bianchi
Additional interferers
B
A D Î case K=4
B C B Ö note that for each
A D A D cluster there are always
B C B C B NI=6 first-ring interferers
D A D A
C B C B C
A D A D
B C B C B
D A D A
C B C B C
A D A D
B C B C B
D A D A
B C B
D A In CCI computation, contribute of
B
additional interferers is marginal
Giuseppe Bianchi
10
sectorization
Î Directional antennas
Î PROS:
Sector 2
Ö CCI reduction
f a , L +1 L f a , 2 L
Î CONS:
Ö Increased handover rate CELL a
Ö Less effective “trunking” leads to
performnce impairments
Giuseppe Bianchi
ÎInferference from 2
cells, only A
G E A
ÖInstead of 6 cells C F G E
D C F
With usual approxs A B D C
(specifically, Dint ~ D) E A B
F G E A
⎡S ⎤ R −η ⎡S ⎤
⎢⎣ I ⎥⎦ o 2 D −η = 3 ⋅ ⎢⎣ I ⎥⎦
= C F G
120 omni B D C F
⎡S ⎤ ⎡S ⎤ A B D
⎢⎣ I ⎥⎦ o dB = ⎢⎣ I ⎥⎦ dB + 4.77 G E A
120 omni F
Conclusion: 3 sectors = 4.77 dB improvement
Giuseppe Bianchi
11
6 sectors
Î60o Directional
antennas
ÎCCI reduction:
Ö1 interfereer only
Ö6 x S/I in the omni case
ÖImprovement: 7.78 dB
Giuseppe Bianchi
PART 2
Cellular Coverage Concepts
Lecture 2.3
teletraffic considerations,
teletraffic planning
Giuseppe Bianchi
12
Traffic generated by one user
(statistical notion of traffic)
traffic)
U2 4
A = ∑ Ai = 4 Ai
i =1
U3
U4
⎛4⎞
P[k active calls] = ⎜⎜ ⎟⎟ Aik (1 − Ai )
4−k
⎝ ⎠
k
E [active calls] = 4 ⋅ Ai = A
TOT
Giuseppe Bianchi
13
example
Î5 users
ÎEach user makes an average of 3 calls
per hour
ÎEach call, in average, lasts for 4 minutes
⎡ calls ⎤ 4
× [hours ] = [erl ]
1
Ai = 3⎢ ⎥
⎣ hour ⎦ 60 5
A = 5 × [erl ] = 1[erl ]
1
5 number of active users
0
probability
0,327680
Meaning: in average, there is 1 active call; 1 0,409600
but the actual number of active calls varies 2
3
0,204800
0,051200
from 0 (no active user) to 5 (all users active), 4 0,006400
with given probability 5 0,000320
Giuseppe Bianchi
Î 30 users
3 4060 1,9E-01 0,863527
4 27405 9,0E-02 0,953564
Î Each user makes an 5
6
142506
593775
3,3E-02
1,0E-02
0,987006
0,996960
average of 1 calls per 7 2035800 2,4E-03 0,999397
hour 8
9
5852925
14307150
5,0E-04
8,7E-05
0,999898
0,999985
Î Each call, in average, 10 30045015 1,3E-05 0,999998
14
A note on binomial coefficient computation
⎛ 60 ⎞ 60!
⎜⎜ ⎟⎟ = = 1.39936e + 12
⎝ ⎠
12 12!48!
but 60!= 8.32099e + 81 (overflow problems! ! )
⎛ 60 ⎞ ⎛ ⎛ 60 ⎞ ⎞
⎜⎜ ⎟⎟ = exp⎜⎜ log⎜⎜ ⎟⎟ ⎟⎟ = exp(log(60!) − log(12!) − log(48!)) =
⎝12 ⎠ ⎝ ⎝12 ⎠ ⎠
⎛ 60 12 48
⎞
= exp⎜ ∑ log(i ) − ∑ log(i ) − ∑ log(i )⎟ (no overflow! ! before exp...)
⎝ i =1 i =1 i =1 ⎠
⎛ 60 ⎞ 12
⎜⎜ ⎟⎟ Ai (1 − Ai ) =
48
⎝ ⎠
12
⎛ 60 12 48
⎞
= exp⎜ ∑ log(i ) − ∑ log(i ) − ∑ log(i ) + 12 log( Ai ) + 48 log(1 − Ai )⎟
⎝ i =1 i =1 i =1 ⎠
(no overflow! ! never! )
Giuseppe Bianchi
Infinite Users
Assume M users, generating an overall traffic intensity A
(i.e. each user generates traffic at intensity Ai =A/M).
We have just found that ⎛ A⎞
M
⎜1 − ⎟
⎛M ⎞
k
M! ⎛ A ⎞ ⎝ M⎠
P[k active calls, M users] = ⎜⎜ ⎟⎟ Aik (1 − Ai )
M −k
= ⎜ ⎟
⎝ ⎠
k ( M − k )!k! ⎝ M ⎠ ⎛ A⎞
k
⎜ 1 − ⎟
Let MÆinfinity, while maintaining the same overall traffic intensity A ⎝ M⎠
M −k
1 Ak ⎛ A⎞ ⎛ A⎞
P[k active calls, ∞ users] = lim
M!
⋅ ⋅ k ⋅ ⎜1 − ⎟ ⋅ ⎜1 − ⎟ =
M → ∞ (M − k )! k! M
⎝ M⎠ ⎝ M⎠
−A
⎡ − ⎤
M
M (M − 1)L (M − k + 1) ⎢⎛
−k
A k
A⎞ A⎥ ⎛ A⎞ Ak
= ⋅ lim ⋅ ⎜1 − ⎟ ⋅ ⎜1 − ⎟ = e− A
k! M → ∞ Mk ⎢⎝ M ⎠ ⎥ ⎝ M⎠ k!
⎣ ⎦
Giuseppe Bianchi
15
Poisson Distribution
30%
poisson
A=2 erl
25% binomial (M=30)
20%
A=10 erl
15%
10%
5%
0%
0 2 4 6 8 10 12 14 16 18 20 22
16
Channel utilization probability
offered traffic: 2 erl - C=3
Î Assumptions: 30%
= C 0%
i =0
Πblock = C C! j = E1,C ( Ao )
blocking probability
10,00%
Ao
∑j =0 j! 1,00%
C=1,2,3,4,5,6,7
0,10%
0,01%
0 1 2 3 4 5
offered load (erlangs)
Giuseppe Bianchi
17
NOTE: finite users
Capacity planning
ÎTarget: support users with a given Grade
Of Service (GOS)
ÖGOS expressed in terms of upper-bound for the blocking
probability
ÆGOS example: subscribers should find a line available in the
99% of the cases, i.e. they should be blocked in no more than
1% of the attempts
ÎGiven:
ÆC channels
ÆOffered load Ao
ÆTarget GOS Btarget
ÖC obtained from numerical inversion of Btarget = E1,C ( Ao )
Giuseppe Bianchi
18
Channel usage efficiency
Offered load (erl) Carried load (erl)
Ao C channels Ac = Ao (1 − B )
Ao B
Blocked traffic
Ac Ao (1 − E1,C ( Ao )) Ao
efficiency : η = = ≈ if small blocking
C C C
Fundamental property: for same GOS, efficiency increases as C grows!!
Giuseppe Bianchi
example
100,0%
A= 40 erl
A= 60 erl
A= 80 erl
blocking probability
A= 100 erl
10,0%
1,0%
0,1%
0 20 40 60 80 100 120
capacity C
19
Erlang B calculation - tables
Giuseppe Bianchi
ÎOnline calculator:
Öhttp://mmc.et.tudelft.nl/~frits/Erlang.htm
Giuseppe Bianchi
20
Application to cellular networks
Cell size (radius R) may be determined
on the basis of traffic considerations
Î First step: Î Third step:
Ö Given num channels and GOS Ö Given density of users
ÆC=50 available channels in a cell Æδ=500 users/km2
ÆBlocking probability<=2% Ö Evaluate cell radius
Ö Evaluate maximum cell (offered) load M M
δ= ⇒ R=
ÆFrom Erlang-B inversion(tables) πR 2
πδ
A=40.25 erl
Ö R~438m
Î Second step
Ö Given traffic generated by each user
ÆEach user: 4 calls/busy-hour
ÆEach call: 2 min in average
ÆAi=4x2/60=0.1333 erl/user
Ö Evaluate max num of users in cell
ÆM=301.87 ~ 302
Giuseppe Bianchi
Other example
Î Three service providers are planning to provide cellular service for an urban
area. The target GOS is 2% blocking. Users make 3 calls/busy-hour, each
lasting 3 minutes in average (Ai=3/20=0.15)
Ö Question: how many users can support each provider?
Î Provider A configuration: 20 cells, each with 40 channels
Î Provider B configuration: 30 cells, each with 30 channels
Î Provider C configuration: 40 cells, each with 20 channels
Î Provider A: Î Provider B: Î Provider C:
Ö 40 channels/cell Ö 30 channels/cell Ö 20 channels/cell
Ö at 2%: Ao=30.99 erl/cell Ö at 2%: Ao=21.93 erl/cell Ö at 2%: Ao=13.18 erl/cell
Ö 619.8 erl-total Ö 654.9 erl-total Ö 527.2 erl-total
Ö M=4132 overall users Ö M=4386 overall users Ö M=3515 overall users
Compare case A with C! The reason is the lower efficiency of 20 channels versus 40
Giuseppe Bianchi
21
Sectorization and traffic
Î Assume cluster K=7
Î Omnidirectional antennas: CCI=18.7 dB
Î 120o sectors: CCI=23.4 dB
Î 60o sectors: CCI=26.4 dB
Giuseppe Bianchi
conclusion
ÎThis module has given some hints regarding:
Ö Cell sizing via propagation considerations
Ö Frequency reuse via propagation considerations
Ö Cell planning via teletraffic consideration
Î Very elementary models
Ö But sufficient to understand what’s inside planning
ÎNo mobility!
Ö Teletraffic models need to be extended to manage handover rates!
Ö Blocking requirement for an handover call MUST be much lower
than blocking for a new incoming call
Æsevere math complications
ÆGuard channels for handover
ÆOut of the scopes of this class!
Giuseppe Bianchi
22
PART 3
Introduction to GSM
Lecture 3.0
History
Giuseppe Bianchi
History of wireless
communication
Î 1896: Marconi
Ö first demonstration of wireless telegraphy
Ö tx of radio waves to a ship at sea 29 km away
Ö long wave transmission, high power req. (200 kW and +)
Î 1901: Marconi
Ö Telegraph across the atlantic ocean
Ö Close to 3000 Km hop!
Î 1907 Commercial transatlantic connections
Ö huge ground stations (30 by100m antennas)
Î 1915: Wireless telephony established
Ö NY – S. Francisco
Ö Virginia and Paris
Î 1920 Marconi:
Ö Discovery of short waves (< 100m)
Ö reflection at the ionosphere
Ö (cheaper) smaller sender and receiver, possible due to the invention of the
vacuum tube (1906, Lee DeForest and Robert von Lieben)
Giuseppe Bianchi
1
History of wireless
communication
Î 1920's: Radio broadcasting became popular
Î 1928: many TV broadcast trials
Î 1930's: TV broadcasting deployment
Î 1946: First public mobile telephone service in US
Ö St. Louis, Missouri
Ö Single cell system
Î 1960's: Bell Labs developed cellular concept
Ö brought mobile telephony to masses
Î 1960’s: Communications satellites launched
Î Late 1970's: technology advances enable
affordable cellular telephony
Ö entering the modern cellular era
Î 1974-1978: First field Trial for Cellular System
Ö AMPS, Chicago
Giuseppe Bianchi
Giuseppe Bianchi
2
1st generation mobile systems
ÎFirst generation: 1980’s ÎAnalog transmission
ÎSeveral competing standards in Ö Frequency modulation
different countries ÎVarious bands:
Ö NMT (Nordic Mobile Telephone) Ö NMT:
ÆScandinavian standard; adopted in most Æ450 MHz first
of Europe Æ900 MHz later
ÆFirst european system (Sweden, 1981) Ö TACS
Ö TACS (Total Access Communication Æ900 MHz
Systems), starts in 1985 Æ1230 bidirectional
ÆUK standard; A few of Europe, Asia, channels (25KHz)
Japan
Ö AMPS
Ö AMPS (Advanced Mobile Phone Æ800 MHz
Service)
ÆUS standard ÎToday still in use in low-
Ö C-Netz (Only in Germany) technology countries
Ö And not yet completely dismissed
Ö Radiocom 2000 (Only in France) in high-tech countries
Giuseppe Bianchi
Giuseppe Bianchi
3
Timing
Î1982: Start of GSM-specification in
Europe (1982-1990)
Î1983: Start of American AMPS
widespread deployment
Î1984 CT-1 standard (Europe) for
cordless telephones
Î1991 Specification of DECT
ÖDigital European Cordless Telephone (today: Digital
Enhanced Cordless Telecommunications)
- ~100-500m range, 120 duplex channels, 1.2Mbit/s data
transmission, voice encryption, authentication
Î1992: Start of GSM operation Europe-
wide
Î1994: DCS-1800
Giuseppe Bianchi
Giuseppe Bianchi
4
3rd generation mobile systems
ÎUMTS (Universal Mobile
TelecommunicationSystem)
ÆITU standard: IMT-2000 (International Mobile
Telecommunication – 2000)
ÆUMTS forum created in 1996
ÆLater on 3GPP forum (bears most of standardization
activities)
ÖWideband CDMA radio interface
ÆBut several other proposals accepted as “compatible”
ÖRadio spectrum: 1885-2025 & 2110-2200 MHz
ÖAlready deployed in Japan
ÖTime to market in Italy: 2004?
Giuseppe Bianchi
Giuseppe Bianchi
5
PART 3
Introduction to GSM
Lecture 3.1
Architecture and components
Giuseppe Bianchi
GSM Network
high-
high-level view
MSC = Mobile Switching Center
= administrative region
PSTN
PSTN
Public
Publicswitched
switched
telephone
telephonenetwork
network
MSC MSC
Base Base
Station Station
PLMN
Public Land
Mobile Network
MSC role: telephone switching central with special mobility management capabilities
Giuseppe Bianchi
6
GSM system hierarchy
MSC MSC region
LOCATION
BSC
AREA MSC: Mobile Switching Center
LA: Location Area
BSC: Base Station Controller
BTS BTS: Base Transceiver Station
BSC
BTS
MS Mobile Station
BTS BTS BTS Base Transceiver Station
BSC Base Station Controller
BTS BSC MSC Mobile Switching Center
GMSC Gateway MSC
OMC Operation and Maintenance Center
BTS
MS EIR Equipment Identity Register
AUC Authentication Center
HLR Home Location Register
VLR Visitor Location Register
Giuseppe Bianchi
7
Î Two components:
GSM Sub-Systems Ö Fixed installed infrastructure
Æ The network in the proper sense
Ö Mobile subscribers
Æ MS: Mobile Station
Î Fixed infrastructure divided
operator into three sub-systems
Ö BSS: Base Station subsystem
Æ Manages transmission path from MS
to NSS
External Ö NSS: Network Switching Subsystem
Networks Æ Communication and interconnection
with other nets
OSS Ö OSS: Operational Subsystem
Æ GSM network administration tools
Users
NSS BSS MS
Giuseppe Bianchi
PART 3
Introduction to GSM
Lecture 3.2
Mobile Station and addresses
Giuseppe Bianchi
8
Mobile Station (MS)
GSM separates user mobility from equipment mobility,
by defining two distinct components
ÎMobile Equipment
ÆThe cellular telephone itself (or the vehicular telephone)
ÖAddress / identifier:
ÆIMEI (International Mobile Equipment Identity)
Giuseppe Bianchi
Terminal
Terminal Mobile
Mobile
Equipment
Equipment Termination
Termination Terminal Terminal Mobile
Equipment Adaptor Termination
9
Mobile Equipment Max Power
5 power classes
I 20 vehicular
II 8 vehicular
III 5 portable
Normally used
IV 2 portable
V 0.8 portable
This was for 900 MHz – for 1800 MHz only two classes: 1W, and 0.25 W
Giuseppe Bianchi
IMEI
International Mobile Equipment Identity
Î Uniquely identifies the mobile equipment
Î 15 digits hierarchical address
Î assigned to ME during manifacturing and “type approval” testing
Ö Type approval procedure: guarantees that the MS meets a
minimum standard, regardless of the manifacturer
Î IMEI structure:
10
IMEI management
Î Protection against stolen and malfunctioning terminals
Î Equipment Identity Register (EIR): 1 DataBase for each operator; keeps:
ÖWHITE LIST:
Ævalid IMEIs
ÆCorresponding MEs may be used in the GSM network
ÖBLACK LIST:
ÆIMEIs of all MEs that must be barred from using the GSM network
ÆException: emergency calls (to a set of emergency numbers)
ÆBlack list periodically exchanged among different operators
ÖGRAY LIST:
ÆIMEIs that correspond to MEs that can be used, but that, for some
reason (malfunctioning, obsolete SW, evaluation terminals, etc), need
to be tracked by the operator
ÆA call from a “gray” IMEI is reported to the operator personnel
Giuseppe Bianchi
SIM card
Subscriber Identity Module
Î Uniquely associated to a user
Ö Not to an equipment, as in first generation cellular networks
Î Stores user addresses
Ö IMSI
Ö MSISDN
Ö Temporary addresses for location, roaming, etc
Î authentication and encryption features
Ö All security features of GSM are stored in the SIM for maximum protection
Æsubscriber’s secret authentication key (Ki)
ÆAuthentication algorithm (“secret” algorithm - A3 – not unique)
ÆCipher key generation algorithm (A8)
Î Personalization
Ö SIM stores user profile (subscribed services)
Ö RAM available for SMS, short numbers, user’s directory, etc
Ö Protection codes
ÆPIN (Personal Identification Number, 4-8 digits)
ÆPUK (PIN Unblocking Key, 8 digits)
Giuseppe Bianchi
11
IMSI
International Mobile Subscriber Identity
Î Uniquely identifies the user (SIM card)
Î GSM-specific address
Ö unlike MSISDN - normal phone number
Î 15 digits hierarchical address
Î assigned by operator to SIM card upon subscription
Î IMSI structure:
MSISDN
Mobile Subscriber ISDN Number
Î MSISDN: the “usual” telephone number
Ö Follows international ISDN numbering plan (ITU-T E.164 recommendations)
Ö Structure:
CC – up to 3 digits NDC – 3 digits (for PLMN) SN – max 10 digits
(Country Code) (National Destination Code) (Subscriber Number)
12
Temporary addresses
Î TMSI - Temporary Mobile Î MSRN - Mobile Station
Subscriber Identity Roaming Number
Ö 32 bits Ö An MSISDN number
Ö assigned by VLR within an ÆCC, NDC of the visited network
administrative area ÆSN assigned by VLR
Æhas significance only in this area Ö Used to route calls to a roaming MS
Ö transmitted on the radio interface ÆSubscriber Number (SN)
instead of IMSI assigned to provide routing
Æ reduces problem of information towards actually
“eavesdropping” responsible MSC
Giuseppe Bianchi
PART 3
Introduction to GSM
Lecture 3.3
Fixed Infrastructure
Giuseppe Bianchi
13
Components and interfaces
Components
MS MS Mobile Station
BTS Base Transceiver Station
Um BSC Base Station Controller
MSC Mobile Switching Center
BTS OMC Operation and Maintenance Center
EIR Equipment Identity Register
Abis AUC Authentication Center
Other HLR Home Location Register
BSC VLR Visitor Location Register
MSC E BSS
A Interfaces
Other
MSC OMC Um Radio Interface
Networks
B Abis BTS-BSC
C A BSS-MSC
F B MSC-VLR
EIR HLR VLR C MSC-VLR
D D HLR-VLR
G E MSC-MSC
F MSC-EIR
Other
AUC G VLR-VLR
VLR
Giuseppe Bianchi
BTS BSC
A
BTS A-bis Interface
Interface
14
Base Transceiver Station - BTS
Um Interface
(to MS)
Transmission
filter Transmitter
Slow freq.
Hopping
TRX
System
Digital
Signal
Input HF Processing
Filter Receiver
Giuseppe Bianchi
15
Base Station Controller - BSC
DB contains
BTS-1 DB - state information for all BSS
- BTS software
BTS-2
X From/to MSC
switch FUNCTIONS:
Î switch calls from MSC to
matrix
correct BTS
Ö and conversely
BTS-K
Î Protocol and coding
conversion
Ö for traffic (voice) &
1 BSC may control signaling (GSM-specific
to ISDN-specific)
up to 40 BTS
Î Manage MS mobility
Î Enforce power control
Giuseppe Bianchi
BTS:
-collects speech traffic
-Deciphers and removes error protection
-Result:
-13 kbps air-interface GSM speech-coded signal Transcoding and
MSC: Rate Adaptation
-A modified ISDN switch Unit (TRAU)
-Needs to receive ISDN-coded speech needed!
-64 kbps PCM format (A-law)
Giuseppe Bianchi
16
TRAU possible placements
64 kbit/s
13 kbit/s (4x16 sub-mux)
On MSC 16 kbit/s
BTS BSC TRAU MSC
Why 16 kbps instead of 13? Inband signalling needed for BTS control of TRAU
(TRAU needs to receive synchro & decoding information from BTS)
Giuseppe Bianchi
Giuseppe Bianchi
17
Mobile Switching Center - MSC
Î An ISDN switch (64 kbps channels)
Î Performs all the switching and routing functions
of a fixed network switching node
Î PLUS specific mobility-related functions:
Ö Allocation and administration of radio resources
Ö Management of mobile users
Æregistration, authentication
Æhandover execution and control
Æpaging
Giuseppe Bianchi
18
Authentication Center - AUC
ÎAssociated to HLR
ÖEventually integrated with HLR
ÎSearch key: IMSI
ÎResponsible of storing security-relevant
subscriber data
ÖSubscriber’s secret key Ki (for authentication)
ÖEncryption key user on the radio channel (Kc)
ÖAlgorithms to compute volatile keys used during
authentication process
Giuseppe Bianchi
PLMN
Public Land
Mobile Network
Giuseppe Bianchi
19
Visitor Location Register - VLR
Î At most 1 database per MSC
Ö Generally, joint MSC-VLR implementation
ÆNo need to carry heavy MSC-VRL signalling load on network links
Ö but 1 VLR may serve many MSCs
Î VLR entries:
Ö Every user / MSISDN actually staying in the administrative area of the
associated MSC
ÆEntry created when an MS enters the MSC area (registration)
Ö NOTE: may store data for roaming users (subscribed to different operators)
Î Stores:
Î Subscriber and subscription data Î Tracking and routing information
Ö IMSI, MSISDN Ö Mobile Station Roaming Number
Ö Parameters for additional services (MSRN)
Ö info about user equipment (IMEI) Ö Temporary Mobile Station Identity
Ö Authentication data (TMSI)
Ö Location Area Identity (LAI) where MS
has registered
ÆUsed for paging and call setup
Giuseppe Bianchi
20
PART 4
GSM – Radio Interface
Lecture 4.1
Physical channels
Giuseppe Bianchi
1
Adjacent channels
(due to GMSK)
60dB 35dB
Specification: 9dB
In practice, due to power control and shadowing, adjacent channels
Cannot be used within the same cell…
Giuseppe Bianchi
Physical channel
Î 200 KHz bandwidth + GMSK modulation
Ö 1625/6 kbps gross channel rate (270.8333 kbps)
Î 1 time slot = 625/4 bits
Ö 156.25 bits
Ö 15/26 ms = 576.9 µs
time time
slot slot
0 7 time
577 µs
Giuseppe Bianchi
2
Hybrid FDMA-TDMA
physical channel = (time slot, frequency)
frequency Total n. of channels: 992
200 KHz
200 KHz
200 KHz
200 KHz
200 KHz
200 KHz slot
200 KHz
200 KHz
200 KHz
577us 577us 577us 577us 577us 577us 577us 577us
time
Giuseppe Bianchi
Giuseppe Bianchi
3
Other GSM bands
ÎExtended GSM (E-GSM) band
ÖUplink: 880-915 MHz
ÖDownlink: 925-960 MHz
ÎOther bands:
Ö450 MHz Æ (450.4-457.6 up; 460.4-467.6 down)
Ö480 MHz Æ (478.8-486 up; 488.8-496 down)
Ö1900 MHz Æ (1850-1910 up; 1930-1990 MHz)
Giuseppe Bianchi
Duplexing
- MS uses SAME slot number on uplink and downlink
0 1 2 3 4 5 6 7 DOWNLINK
UPLINK 0 1 2 3 4 5 6 7
Giuseppe Bianchi
4
Slow Frequency hopping
(optional procedure within individual cell)
f7
f6
f5
f4
f3
f2
f1
Hopping sequence (example):
… Æ f1 Æ f2 Æ f5 Æ f6 Æ f3 Æ f7 Æ f4 Æ f1…
Î Symmetric structure
Î DATA: 2 x 57 data bits
Ö 114 data bits per burst
Ö “gross” bits (error-protected; channel coded)
Ö “gross” rate: 24 traffic burst every 26 frames (120 ms)
Æ22.8 kbps gross rate
Æ13 kbps net rate!
Î S: 2 x 1 stealing bit
Ö Also called stealing flags, toggle bits
Ö Needed to grab slot for FACCH (other signalling possible)
Giuseppe Bianchi
5
Tail & training bits
Î 2 x TB = 3 tail bits set to 000
Ö At start and end of frame
Ö Leave time available for transmission power ramp-up/down
Ö Assures that Viterbi decoding starts and ends at known state
Giuseppe Bianchi
Training sequences
Giuseppe Bianchi
6
Power mask for Normal Burst
7
Guard period sizing
dwlink slot 1 … dwlink slot 4
… uplink slot 1
Frame synchronization
TA (transmitted in the SACCH)
BTS
8
Timing Advance analysis
Î Downlink propagation delay:
Ö d/c
Î Uplink propagation delay:
Ö d/c
Î Uplink delay with TA:
Ö d/c-TA
Î Perfect resynchronization occurs when
Ö TA = 2d/c
Î Maximum cell size for perfect resync:
31.5 [bits ]
⋅ 300000 [km / s ] = 34.89 [km]
TA
d= ⋅c =
2 270833 [bits / s ]
8.25 bits Guard time additionally available for imperfect sync (+/- error)
Giuseppe Bianchi
Access burst
Training
TB Data TB GP
sequence
8 41 36 3 68.25
88 bit burst
156.25 bit (0.577 ms)
Giuseppe Bianchi
9
Other burst formats in GSM
Î5 different bursts available
ÖNormal Burst
ÖAccess Burst
Giuseppe Bianchi
Giuseppe Bianchi
10
Dummy Burst
Dummy Burst
Training
TB Fixed bit pattern Fixed bit pattern TB GP
sequence
3 58 58 3 8.25
26
Giuseppe Bianchi
Synchronization Burst
Synchronization Burst
Training
TB Sync data Sync data TB GP
sequence
3 39 39 3 8.25
64
11
PART 4
GSM – Radio Interface
Lecture 4.2
logical channels
Giuseppe Bianchi
Logical channels
(traffic channels, signalling (=control) channels)
Physical channels
(FDMA/TDMA)
Î Physical channels
Ö Time slots @ given frequencies
Ö Issues: modulation, slot synchronization, multiple access techniques,
duplexing, frequency hopping, etc
Î Logical channels
Ö Built on top of phy channels
Ö Issue: which information is exchanged between MS and BSS
Giuseppe Bianchi
12
Logical – physical mapping
Physical Channel: data rate r, time slot i
frequency
frequency
Frame 8 Frame 9 Frame 10 Frame 11 Frame 12
Logical channel A: data rate r/3, time slot i, frame 3k
Logical channel B: data rate 2r/3, time slot i, frame 3k+1, 3k+2
Giuseppe Bianchi
13
An example procedure involving signalling
Steps:
- paging for MS
- MS responds on RACH
- MS granted an SDCCH
- authentication & ciphering on SDCCH
- MS granted a TS (TCH/FACCH)
- connection completed on FACCH
- Data transmitted on TCH
Giuseppe Bianchi
PART 4
GSM – Radio Interface
Lecture 4.3
Traffic channels and
associated signalling channels
Giuseppe Bianchi
14
Traffic channels (TCH/F)
Periodic pattern of 26 frames (120 ms = 15/26 ms/TS * 8 TS/frame* 26 frame)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Speech coding
160 voice
samples Speech CODER 260 bits block
(20 ms) (8:1 compression) 13 kbps
(2080 bits)
Giuseppe Bianchi
15
Discontinuous transmission
Î Speech coder implements Voice Activity Detection (VAD)
Ö Voice activity: idle for about 40% of the time
Ö To avoid clipping: hangover period (80ms)
Giuseppe Bianchi
Channel Coding
Coding: needed to move from 10-1 to 10-3 radio channel native BER
down to acceptable range (10-5 to 10-6) BER
260 bits
260 bits block divided into
182 bits 78 bits -Class I: important bits (182)
-Class Ia: Most important 50
Parity bits
Tail bits(0000) -Class Ib: Less important 132
-Class II: low importance bits (78)
50 bits 3 132 bits 4
456 bits
First step: block coding for error detection in class Ia (errorÆdiscard frame)
Second step: convolutional coding for error correction
Giuseppe Bianchi
16
Block interleaving
8
0 1 2 3 4 5 6 7
8 9 10 11 12 13 14 15
16 17 18 19 20 21 22 23
.. .. .. .. .. .. .. ..
57
.. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. ..
448 449 450 451 452 453 454 455
B1 B2 B3 B4 B5 B6 B7 B8
Giuseppe Bianchi
Diagonal Interleaving
Block n-1 Block n Block n+1
B1 B2 B3 B4 B5 B6 B7 B8 B1 B2 B3 B4 B5 B6 B7 B8 B1 B2 B3 B4 B5 B6 B7 B8
Inter-burst Interleaving
= Bxn −1 = Bxn− 4
Training
TB S S TB GP
sequence
3 1 26 1 3 8.25
Giuseppe Bianchi
17
Bad Frame Indicator
Giuseppe Bianchi
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Giuseppe Bianchi
18
SACCH block
Î184 signalling bits
ÆBlock coding adds 40 bits (=224)
Æ4 tail (zero) bits (=228)
Æ1/2 Convolutional encoding (=456 bits)
ÎInterleaving: B1 B2 B3 B4 B5 B6 B7 B8
Ö 8 blocks of 57 bits;
Ö spreaded into four consecutive bursts
Ö4 bursts = 1 and only 1 SACCH block!
ÎSACCH rate: Odd/even
Ö184 bits/480 ms = 383.3 bit/s interleaving
SACCH contents
Î 184 bits = 23 bytes
0 1 2 3 4 5 6 7
Î Power level
free Power level
Î Timing advance
free Timing advance
Î Measurement reports for
link quality
Î Measurement reports for (21 bytes – datalink layer)
handover management Includes measurement reports
Giuseppe Bianchi
19
Power control
Î MS has ability to reduce/increase
Maximum power power
(defined by class) Ö Up to its power class maximum
Î Maximum one 2 dB step every 60
ms
Î uplink power measures taken by
BTS
2 dB steps;Î notified back to MS via SACCH
Ö Power level values: 0-15
Æ0 = 43 dBm (20 W)
Æ15 = 13 dBm (20 mW)
Î algorithm: manifacturer specific
Ö runs on BSC
Minimum power
(13 dBm for GSM) Î power control applied also on
(0 dBm for DCS 1800) downlink
Giuseppe Bianchi
Measurement values
ÎRXLEV ÎRXQUAL
Ö Power level Ö Bit Error Rate (raw)
Ö Present channel + neighbohr cell)
Bit error From To
RX signal From To Ratio (%) (%)
level (dBm) (dBm) RXQUAL_0 - 0.2
RXLEV_0 - -110 RXQUAL_1 0.2 0.4
RXLEV_1 -110 -109 RXQUAL_2 0.4 0.8
RXLEV_2 -109 -108
RXQUAL_3 0.8 1.6
RXLEV_3 -108 -107
… … … RXQUAL_4 1.6 3.2
20
Fast Associated Control Channel
Î FACCH: urgent signalling
Ö Used when several signalling information needs to be transmitted
ÆCall setup
ÆHandover
Î FACCH block = 184
Ö 456 after coding
Î Interleaved as voice block
Ö Spreaded on 8 bursts
Î Replaces a voice block (20 ms) on the TCH
Ö Via stealing bits
Ö Voice block(s) discarded
Î Maximum FACCH bit rate
Ö 184*6/120 [bits/ms] = 9.2 kbps (vs 383 bps of SACCH!)
Giuseppe Bianchi
time
Giuseppe Bianchi
21
Half-rate traffic channels (TCH/H)
ÎSupport for 5.6 kbps voice codecs
ÎSpecification in 1995
Î228 bits block
Ö112 bits of compressed 20ms voice
Æ95 bits class I
» + 3 parity + 6 tail + convolutional coding 104/211
Æ17 bits class 2
ÎInterleaving:
ÖSame as voice (block + diagonal + odd/even)
ÖBut on 4 bursts
ÎFraming:
TCH/H 0…7 TCH/H 8…15
[subchannel 0] [subchannel 1] SACCH 0…7 SACCH 8…15
1 2 3 4 5 6 7 8 9 10 11 12
Giuseppe Bianchi
ÎTCH data
ÖBasic speed: 9.6 kbps (data,fax)
ÖOther speeds: <2.4; 4.8; 9.6; 14.4;
ÎDifferent coding details
ÖSee text(s)
ÎMajor difference with voice:
ÖInterleaving with depth = 19 (!!!)
Öcomplete fading of a burst is recoverable (unlike
voice)
Giuseppe Bianchi
22
PART 4
GSM – Radio Interface
Lecture 4.4
Broadcast Carrier and
Channels
Giuseppe Bianchi
10 frame sub-block
Giuseppe Bianchi
23
BCCH carrier placement
ÎOn Downlink
ÖCorresponding uplink dedicated to Random Access Channel
51 frame structure - uplink
RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR
Giuseppe Bianchi
MS powering up
Giuseppe Bianchi
24
tuning
ÎMS listens on strongest beacon for a pure
sine wave (FCCH)
ÖCoarse bit synchronization
ÖFine tuning of oscillator
ÎImmediately follows SCH burst
ÖFine tuning of synchronization (64 bits training sequence)
ÖRead burst content for synchronization data
Æ25 bits (+ 10 parity + 4 tail + ½ convolutional coding = 78
bits)
Æ6 bits: BSIC
Æ19 bits: Frame Number (reduced)
ÎFinally, MS can read BCCH
Giuseppe Bianchi
Multi-framing structure
8 TS, 4.615 ms
0 1 2 3 4 5 6 7
1 multiframe = 26 TDMA frames (120 ms) 1 multiframe = 51 TDMA frames (235.38 ms)
Giuseppe Bianchi
25
Why frame number?
ÎFrame #
ÖDistinguishes logical channels in the same physical channel
ÎMultiframe #
ÖDetermines how BCCH is constructed
ÆI.e. which specific information transmitted on BCCH during a
given multiframe
ÎSuperframe #
ÖUsed as input parameter by encyphering algorithm
Giuseppe Bianchi
BCCH contents
Î 184 bits
Ö Coded in 456 bits and interleafed in 4 bursts
Ö same coding and interleaving as SACCH
Ö BCCH capacity
Æ184 bits / (51*8*15/26 ms) ~ 782 bps
Î Information provided
Ö Details of the control channel configuration
Ö Parameters to be used in the cell
ÆRandom access backoff values
ÆMaximum power an MS may access (MS_TXPWR_MAX_CCCH)
ÆMinimum received power at MS (RXLEV_ACCESS_MIN)
ÆIs cell allowed? (CELL_BAR_ACCESS)
ÆEtc.
Ö List of carriers used in the cell
ÆNeeded if frequency hopping is applied
Ö List of BCCH carriers and BSIC of neighboring cells
Giuseppe Bianchi
26
control channel alternative
organization
51 frame structure – small capacity cell
DOWN FSBBBBPPPPFSPPPPPPPPFSDDDDDDDDFSDDDDDDDDFSDDDDDDDD
UP DDDDRRDDDDDDDDRRRRRRRRRRRRRRRRRRRRRRRDDDDDDDDRRDDDD
UP RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR
TTTTTTTTTTTT S TTTTTTTTTTTT
- important slots while call is active: frequency correction FCCH and sync SCH!
- needed for handover
-Worst case: at most every 11 TCH multiframes (1.32 s), there will be
a frequency correction burst of a neighboring cell
Giuseppe Bianchi
27
PART 4
GSM – Radio Interface
Lecture 4.5
Paging, Random Access,
dedicated signaling
Giuseppe Bianchi
Why paging
ÖChannel assignment:
Æonly upon explicit request from MS
ÖPaging
Æneeded to “wake-up” MS from IDLE state when incoming call
arrives to MS
ÖMS accesses on RACH to ask for a channel
ÆGenerally SDCCH (but immediate TCH assignment is possible)
1) paging
3) Channel assignment
28
Paging
ÎPaging message generated by MSC
ÖWhich receives incoming call
ÎTransferred to subset of BSC
ÖPaging limited to user’s location area
ÖPaging message contains:
ÆList of cells where paging should be performed
ÆIdentity of paged user (IMSI or TMSI)
ÎPaging message coded in 4 consecutive
bursts over the air interface
ÖSame coding/interleaving structure of SACCH (184Æ456 bits)
ÎPaging for more MSs may be joined in one
unique paging message
Giuseppe Bianchi
Giuseppe Bianchi
29
Access procedure
Î Always activated by MS
Ö In response to paging (incoming call)
Ö When location update needs to be performed
Ö When new call is generated by MS
Î Based on access burst sent on RACH
Ö Access burst coding: 8 payload bits (channel_req) Æ 36 coded bits
Æ+6 parity; + 4 tail; + ½ convolutional coding
Access burst
TB Training Data TB GP
8 Seq. (41) 36 3 68.25
RACH operation
MS-C
MS-A MS-B
Multiple Access Technique for simultaneous access
Collision resolution based on
- random retrial period
- “permission” probability Same thing..!
(SLOTTED ALOHA protocol)
Giuseppe Bianchi
30
RACH performance
ÎN stations
ÎEach transmits with probability p
ÖRetries, in average, every 1/p slots
Îrelevant probabilities:
idle : (1 − p )N
Np (1 − p )
N −1
success :
N
⎛N⎞
∑ ⎜⎜ k ⎟⎟ p (1 − p )
k N −k
collision :
k =2 ⎝ ⎠
Maximum efficiency: when p=1/N
Giuseppe Bianchi
Giuseppe Bianchi
31
Access signaling - 1
MS BTS BSC MSC
Channel_request
rnd number
Channel_required
rnd&frame number,
Delay (TA estimate)
Channel_activation
Ch_activation_ack
Immediate_assignment
rnd&frame number, channel description,
Initial TA, initial max power
Giuseppe Bianchi
Giuseppe Bianchi
32
If same random discriminator?
Î two MSs may have same random discriminator
Ö Likely in heavy load, with only 5 bits
Î And transmit in the same frame
Î Only one wins (the other is faded)
Î contention resolved via explicit MS identification
Ö On SDCCH MS1 MS2 BTS
ID1
ID2
ID1
leave
continues
Giuseppe Bianchi
Access signaling - 2
MS BTS BSC MSC
… … … …
Immediate_assignment
Initial_message
Giuseppe Bianchi
33
Stand-alone Dedicated Control
Channel - SDCCH
Î Dedicated bidirectional channel to MS
Ö but used to exchange signalling
Î Has an associated SACCH
Î Coding: as SACCH
Ö 184Æ456 bits on 4 consecutive bursts
SDCCH/4 for small
Î Typical framing (SDCCH/8)
cells – SDCCH shares
Ö 8 SDCCH (+8 SACCH) on 1 channel (carrier,TS) BCCH+PAGCH channel
Ö 1 SDCCH message per 51-multiframe - see before -
Æ184 bits / (51*8*15/26 ms) = 598/765 kbps ~ 782 bit/s
Ö 1 SACCH every 2 multiframe
DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDAAAAAAAAAAAAAAAA
000011112222333344445555666677770000111122223333
DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDAAAAAAAAAAAAAAAA
000011112222333344445555666677774444555566667777
Giuseppe Bianchi
34
PART 5
GSM – Switching & Mobility
Lecture 5.1
Protocol architecture overview
Giuseppe Bianchi
Giuseppe Bianchi
1
RR
Î Manages administration of frequencies and channels
Ö Mostly deals with air interface
ÆSeveral RR functions considered in previous part
Î Guarantees stable link upon handover
ÆSurprise! handover is part of RR, not MM!
Î Function summary:
Ö Monitoring BCCH, PCH
Ö RACH administration
Ö Request/assignment of channels
Ö MS power control & synchronization
Ö Handover
Î Where is RR:
Ö MS, BTS, BSC, MSC
Giuseppe Bianchi
MM
Î Manages user location and tasks resulting from
mobility
Î Function summary:
Ö TMSI assignment
Ö MS localization
Ö Location updating
Ö MS authentication
Ö MS identification, attach/detach
Î Where is MM:
Ö MS, MSC
Giuseppe Bianchi
2
CM
Î Controls calls, supplementary services, and SMS
Î Function summary:
Ö Call establishment (from MS, to MS)
Ö Emergency call management
Ö Call termination
Ö DTMF signaling (Dual Tone MultiFrequency)
Ö In-call modification
Î Where is CM:
Ö MS, MSC, GMSC
Giuseppe Bianchi
Protocol placement
CM
GMSC
MM
HLR
RR
Trans.
MSC
MS BTS BSC
(VLR)
Giuseppe Bianchi
3
Protocol outline
Relay Anchor
MS BTS BSC MSC MSC HLR
CM RIL3-CC
MAP/D
MM RIL3-MM
RR RIL3-RR
RSM BSSMAP MAP/E
TCAP
PART 5
GSM – Switching & Mobility
Lecture 5.2
handover (physical mobility)
Giuseppe Bianchi
4
Neighbor cells
BTS
2
BTS BTS
1 n
Î A station must:
Ö monitor beacon power level of neighbor cells
Ö Keep detailed track of best 6 neighbor cells
Ö DECODE their BCCH (i.e. read FCCH, SCH) to get parameters
ÆAt least once every 5 minutes
ÆBSIC (from SCH) refreshed every at most 30s
Giuseppe Bianchi
When cell parameters are the same, simply select cell with higher RXLEV!
Giuseppe Bianchi
5
Cell reselection criterion (C2)
ÎReselect cell with greatest C2>0:
Giuseppe Bianchi
BTS BTS
6
handover
Î Procedure in which an MS releases a connection with
a BTS, and establishes a connection with a new BTS,
while ensuring that the ongoing call is maintained
Ö The MS remains in dedicated state (unlike cell reselection, where MS is in idle
state)
Î Handoff: synonymous of handover
Î Needs two mechanisms
Ö Handover preparation: detection of cell-border crossing
ÆBased on radio link quality measurements
Ö Handover execution: setup of a new channel in a cell, and tear-down of a
previous channel
Î Improved handover mechanisms:
Ö Seamless handover: when active call performance is not impaired
ÆNot possible in GSM: for about 100-200ms, communication is interrupted
Ö Soft Handover: when two channels are simultaneously set-up (old and new)
ÆNot possible in GSM; possible in UMTS
Giuseppe Bianchi
Seamless
BSS 1 BSS 2 BSS 1 BSS 2 BSS 1 BSS 2
(DECT)
f1 f1 f2 f2
MS MS MS
Giuseppe Bianchi
7
Handover classification
Classification by motivation Classification by typology
Types of handover
Anchor MSC: Relay MSC:
the MSC that first the MSC that currently
managed the current call manages the call
Switching Switching
point for point for
all inter-MSC A-MSC R-MSC
inter-BSC
handover handover
A
Switching
point for
internal BSC BSC BSC
handover
A-bis
Giuseppe Bianchi
8
Handover taxonomy
Î BCHO: Base station Controlled Handover
Ö Handover detection: BS
Ö Handover Execution: BS
Î MCHO: Mobile Controlled Handover
Ö Handover detection: MS
Ö Handover Execution: MS
Î MAHO: Mobile Assisted Handover
Ö Handover detection: MS
Ö Handover Execution: BS
Giuseppe Bianchi
Ö Down-link signal quality (BER) received from serving cell RXQUAL_2 0.4 0.8
Ö Down-link signal level received from n-th neighbor cell RXQUAL_4 1.6 3.2
RXQUAL_7 12.8 -
Giuseppe Bianchi
9
A note on MS distance
TA[bits ]⋅ tbit =
2d TA
⇒ d= ⋅ c ⋅ tbit
c 2
ÖHence, the TA resolution, in mt, is:
1
300000[mt / ms] ⋅ [ms ]
c ⋅ tbit
d (TA) = TA = TA 270.833 ≈ TA ⋅ 554mt
2 2
ÖINSUFFICIENT for microcells!
ÖSufficient only to understand we are going out of the cell
Giuseppe Bianchi
Handover preparation –
additional metrics
Î Transmission power
Ö Maximum MS transmission power
RXLEV RXLEV
Ö Maximum serving BTS transmission power (cell A) Handover (cell B)
Ö Maximum neighboring BTSs transmission
power
Î congestion status
Ö of serving BTS
Ö of neighboring BTSs
Æ provided they can support the MS. RXLEV RXLEV
Î Handover Margin (cell A)
Handover
(cell B)
Ö To avoid ping-pong handover effect
Ö 5-10 dB in normal operation; up to 30dB in
urban operation (to fight shadowing)
Giuseppe Bianchi
10
handover procedure skeleton
1) Handover request goes up to switching point
MSC
2) Switching point prepares new path on fixed net
Giuseppe Bianchi
Giuseppe Bianchi
11
Inter-MSC handover
Î More complex, as an ISDN circuit must be set
between MSCs
Ö We’ll not enter into details (just the basic ideas)
Î Two cases
Giuseppe Bianchi
PART 5
GSM – Switching & Mobility
Lecture 5.3
location registration/update
Authentication & Ciphering
Giuseppe Bianchi
12
Location Area
vs
MSC service area
Giuseppe Bianchi
Registration vs update
Î Very similar procedures, with goals:
Ö Determine where the user is
Ö Authenticate user
Î Differences:
Ö Location Registration
ÆUser first access to PLMN
» Needs to send IMSI and receive TMSI
Ö Location Update
ÆSubsequent accesses to PLMN (either in old or new MSC/VLS)
» Also after MS shut-down!
» TMSI-based identification
Î Registered user:
Ö The PLMN knows the LA where the user is (or is supposed to be)
Giuseppe Bianchi
13
Procedure start-up
ÎMS switches on
ÎDetects BCCH carrier
ÖTune and synchronize
ÎListens to BCCH
ÎObtains Location Area Identifier
ÖLAI: [CC,MNC,LAC]
ÆCountry Code (CC): 3 digits
ÆMobile Network Code: 2 digits
ÆLocation Area Code: max 5 digits
Giuseppe Bianchi
3
MSC VLR HLR
BSC
BTS
2 1) Obtain LAI from BCCH
BTS
2) Register MS ID into local VLR
1 BTS
3) Update pointer at HLR
MS
Giuseppe Bianchi
14
Location Registration
MS BSS/MSC VLR HLR AUC
Loc. Upd. Request
Update Loc. Area
IMSI, LAI Auth. Param. Req. Auth. Info. Req.
IMSI, LAI
IMSI IMSI
Auth. Info Auth. Info
(Auth. Parameters) (Auth. Parameters)
authentication
Update Location
Start Ciphering IMSI, MSRN
Kc Insert Subscrib. Data
Activate Forward new TMSI IMSI, additional data
ciphering TMSI Insert Subscrib. Data
ACK
Locat. Upd. Accept
Locat. Upd. Accept
IMSI
Authentication
(managed by VLR)
A3 SRES
SRES
Authentication Response
Equal?
Signed RESult: 32 bit SRES
Ki RAND
A8
Kc
Giuseppe Bianchi
15
Authentication (details)
Î Side effect of authentication:
Ö Generate encryption key Kc via A8 algorithm
Î Secret A3, A8 algorithms (one-way hash functions)
Ö Stored into the SIM
ÆAlong with secret key Ki
Ö Note that roaming operator DOES NOT need to know them!
ÆSince A3,A8 run ONLY in the AUC at the home HLR
ÆKi is NEVER transmitted away from AUC or MS!
Î Generally implemented together
Ö [SRES,Kc] = A38[Ki,RAND]
Î To reduce signaling, real implementation slightly
different:
Ö VLR sends IMSI
Ö Receives back several tuples of (RAND, SRES, Kc) to be used for the
considered MS also in subsequent accesses
Giuseppe Bianchi
ciphering
Î A5 algorithm is known (to allow roaming)
Î Generates two ciphering sequences
Ö one for uplink, one for downlink
Ö Sequence periodic with period 26x51x2048=2,715,648
Æ 221=2,097,152 < 2,715,648 < 222=4,194,304
Î 114 bits per frame, depending on frame number
Î XOR-ed with burst data field
MS A5 A5 BTS
S1 S2 S1 S2
Giuseppe Bianchi
16
Location Update in same VLR
(same as location registration, but with TMSI)
MS BSS/MSC VLR HLR AUC
Loc. Upd. Request
Update Loc. Area
TMSI, LAI
TMSI, LAI
authentication Update Location
IMSI, MSRN
Generate
New TMSI
Start Ciphering
Activate Kc Insert Subscrib. Data
ciphering Forward new TMSI IMSI, additional data
Ins. subs. data ACK
Locat. Upd. Accept
Locat. Upd. Accept
IMSI
TMSI Realloc Cmd
Locat. Upd. Accept
TMSI Realloc ACK TMSI ACK
Auth. Param. Req. Auth. Info. Req.
IMSI IMSI
Auth. Info Auth. Info
N x (Kc,RAND,SRES) N x (Kc,RAND,SRES)
Giuseppe Bianchi
Changing MSC/VLR
HLR
VLR MSC Public
Publicswitched
switched
telephone
telephonenetwork
network
PSTN
PSTN
VLR MSC
Base
Station
Base
Station
17
TMSI
ÎTMSI = Temporary Mobile Subscriber Identity
Ö4 octets (32 bits) Operator may set a 6min
ÖRenewed periodically; at every LU / IMSI_attach up to 24hrs periodicity
for LU (value transmitted
ÆVia TMSI_Reallocation_Command/TMSI_Reallocation_Complete on BCCH)
ÆRATIONALE: renew TMSI when transmitted in clear!
IMSI_attach = a special LU
(TMSI reallocation occurs in ciphering mode)
in a same Location Area;
ÎMeaningful only in a given VLR IMSI_attach follows
ÖSpecifically, only for a given Location Area!! an IMSI_detach
(power-down of MS)
ÆSome author (Mouly-Pautet) uses the term
» TIC (Temporary Identity Code) = 4 bytes
» TMSI = TIC+LAI = unambiguous user identification
ÎWhile entering a new Location Area:
Ö user must identify itself with TMSI+LAI pair.
Giuseppe Bianchi
Giuseppe Bianchi
18
Special cases
1. New VLR not capable of determining old
VLR from old LAI
2. Old VLR does not recognize TMSI
MS MSC
Identity Request
Ö Identification procedure
Æ IMSI transmitted in clear Identity Response
IMSI
PAGING:
- Normally based on TMSI
- But when no valid TMSI information available (e.g. after a DB restore
after crash), based on IMSI
Giuseppe Bianchi
PART 5
GSM – Switching & Mobility
Lecture 5.4
Call Management & routing
Giuseppe Bianchi
19
Notation
ÎA call involves two “Parties”
ÎCalling Party (caller)
Öuser generating the call
ÎCalled Party (callee)
Öuser receiving the call
ÎMobile Originating Call (MOC)
ÖCall originated by an MS
ÎMobile Terminating Call (MTC)
ÖCall directed to an MS
Giuseppe Bianchi
DATA DATA
Giuseppe Bianchi
20
Call establishment steps
MS Mobile network MS Mobile network
Terminated Call Originated Call
Paging request
Channel request Channel request
Immediate Assignment Immediate Assignment
Paging Response Service Request
Authentication Request Authentication Request
Authentication Response Authentication Response
Ciphering mode command Ciphering mode command
Ciphering Mode Complete Ciphering Mode Complete
Setup Setup
Call Confirmed Call proceeding
Assignment Command Assignment Command
Assignment Complete Assignment Complete
Alerting Alerting
Connect Connect
Connect Acknowledge Connect Acknowledge
Giuseppe Bianchi
21
DTMF signaling
Î Dual-Tone Multi-Frequency On air interface:
Ö Digital tones associated to terminal keys Ö Signal trasmitted on FACCH as
Æ‘0’…’9’…’#’… signalling data (code of pressed
Î Inband signalling key)
Ö transmitted in the traffic channels! Ö Otherwise coded compression
would distort DTMF tones
Ö Not in the signalling network
Ö Tone generated at MSC when
STOP DMTF message received
MS MSC
Start DTMF (w. key code)
On FACCH)
Key Start DTMF ACK
Pressed
Stop DTMF
Giuseppe Bianchi
Routing an MTC
ISDN
N
SD
SI
M
1:
PLMN
GMSC
2: 3: M
M
N
MSCC
SI SRN
SR
SD
M
N
4:
5: MSRN
VLRB MSCB HLR
6: TMSI MSCA
7:
pa
gi
ng
Giuseppe Bianchi
22
Routing an MTC (alternative)
reduces signalling load during LU
N
D
MSRN retrieved on a per-call basis!
S
SI
(choice of solution depends on trade-offs)
M
1:
PLMN
GMSC
2: 5: M
M
N
MSCC
SI SRN
SR
SD
M
N
6:
7: MSRN
VLRB MSCB HLR
8: TMSI MSCA
9:
pa
gi
3: IMSI
ng
4: MSRN
Giuseppe Bianchi
International
Switching
ISDN Center
(ita)
Transit
Exchange
GMSC 1 PLMN 1
(ITA)
Local MSISDN MSRN MSC
Exchange +39.335.1234567 +44.NDC.8877665
HLR
335.1234567
Giuseppe Bianchi
23
“tromboning”
Call to MSISDN
+39.335.3043125
MSC
ISC
(UK) PLMN 2
(UK)
Call to MSISDN
+39.335.1234567
HLR
Giuseppe Bianchi
Giuseppe Bianchi
24
RLC at ISC - Location Registration
and call management
PLMN 2
ISC
RLC 3 bis (UK) 2 VLR MSC (UK)
PLMN 1
ISC (ITA)
(ITA) 4 HLR
Giuseppe Bianchi
Giuseppe Bianchi
25
SMS routing management
MSC PLMN
Internet, PSDN
IWMSC
Short Message
Get routing info
Service Center for terminating MS
SMS-GMSC HLR
PLMN
MSC
Giuseppe Bianchi
Protocol hierarchy
MS
SM-SC
Short Message
Application Layer
(SM-AL)
Short Message Transfer Protocol (SM-TP)
Short Message
Transfer Layer
(SM-TL) MSC IW-MSC
Short Message Short Message Short Message
Relay Layer Relay Entity Short Message Relay Entity
(SM-RL) (SMR) Relay Protocol (SMR)
(SM-RP)
Connection Short Message Short Message Short Message
Management Control Entity Control Protocol Control Entity
Sublayer (CM-sub) (SMC) (SM-CP) (SMC)
Giuseppe Bianchi
26
Number portability
ÎSubscriber may switch operator without
changing his number
ÎFirst implemented in fixed network
ÖRecently (may 2002) extended to mobile networks
ÎEssential for fair competition among
network operators
ÖUK survey: 42% of corporate subscribers were willing to
change mobile operator; but 96% were, if number could be
ported
ÎResistence from leading operators
ÖNumber portability helps newer operators to compete with
traditional ones
Giuseppe Bianchi
Notation
ÎDonor switch
ÖThe switch corresponding to a “ported” telephone
number
ÎRecipient switch
ÖThe switch to which the ported number is attached
Giuseppe Bianchi
27
Technical solutions
a) call forwarding
Originating network Donor network
switch switch
switch
Recipient network
Originating switch sets-up trunk to donor switch
Donor switch sets-up trunk to recipient switch
Simplest solution, as call forwarding is a feature available in virtually all switches
But extremely inefficient routing and trunking resource consumption!
Giuseppe Bianchi
Technical solutions
b) query on release
Originating network Donor network
SS7 ISUP IAM
switch switch
SS7 ISUP REL
Number
Portability switch
DataBase
Recipient network
Giuseppe Bianchi
28
Technical solutions
c) all-call query
Originating network Donor network
switch switch
Number
Portability switch
DataBase
Recipient network
Giuseppe Bianchi
29
Mobile Number Portability
(with all call query approach)
approach)
Recipient network
Query MSRN
MSC GMSC Return MSRN HLR
IRN
Incoming call
Query IRN Number
switch Return IRN Portability
DataBase
Giuseppe Bianchi
Recipient network
MSRN
Signaling relay
function
Incoming call
Query MSRN Number
switch Return MSRN Portability
DataBase
Giuseppe Bianchi
30